Adi Robertson

We expect that, at least while the appeal is pending, there will be changes to our lending program, but the full scope of those changes is a question pending with the district court. We will provide an update on those changes once the district court decision is final.

Our fight is far from over — We remain steadfast in our belief that libraries should be able to own, preserve, and lend digital books outside of the confines of temporary licensed access. We believe that the judge made errors of law and fact in the decision, and we will appeal.

[F]or libraries in a resource-sharing ‘consortium’ — which seems to be most small and mid-sized public libraries, as well as some large ones — the new system is functionally useless. It’s no longer clear which titles were requested by their own patrons, as opposed to the patrons of another library in the same consortium.

In order to get that information — which until May was freely available — each library now has to pay a separate fee for an “Advantage” account.

In my opinion, the most interesting creative use of large language models is to generate text that’s nothing like a human would have written. [...] In that sense, BLOOM, with its less-perfect retrieval of human output, is better at this task than GPT-4.

It is creepy to me however that the only reason this method gets BLOOM to generate weird designs is because I spent years seeding internet training data with lists of weird AI-generated text.

Today we are launching our government server on Mastodon.

In the letter I sent to the House on ‘Digital Community Goods’ on 22 June, I indicated that I support two value-driven alternatives to social media.

One of these alternatives is a ‘private’ Mastodon server used by government organisations.

With this, we are taking another step in our Value-driven Digitization Work Agenda.